Dual Compensation Residual Networks for Class Imbalanced Learning

doi:10.1109/TPAMI.2023.3275585

	Dual Compensation Residual Networks for Class Imbalanced Learning
	Hou, Ruibing 1,2; Chang, Hong 1,2; Ma, Bingpeng 3; Shan, Shiguang 1,2; Chen, Xilin 1,2
	2023-10-01
发表期刊	IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
ISSN	0162-8828
卷号	45 期号:10 页码:11733-11752
摘要	Learning generalizable representation and classifier for class-imbalanced data is challenging for data-driven deep models. Most studies attempt to re-balance the data distribution, which is prone to overfitting on tail classes and underfitting on head classes. In this work, we propose Dual Compensation Residual Networks to better fit both tail and head classes. First, we propose dual Feature Compensation Module (FCM) and Logit Compensation Module (LCM) to alleviate the overfitting issue. The design of these two modules is based on the observation: an important factor causing overfitting is that there is severe feature drift between training and test data on tail classes. In details, the test features of a tail category tend to drift towards feature cloud of multiple similar head categories. So FCM estimates a multi-mode feature drift direction for each tail category and compensate for it. Furthermore, LCM translates the deterministic feature drift vector estimated by FCM along intra-class variations, so as to cover a larger effective compensation space, thereby better fitting the test features. Second, we propose a Residual Balanced Multi-Proxies Classifier (RBMC) to alleviate the under-fitting issue. Motivated by the observation that re-balancing strategy hinders the classifier from learning sufficient head knowledge and eventually causes underfitting, RBMC utilizes uniform learning with a residual path to facilitate classifier learning. Comprehensive experiments on Long-tailed and Class-Incremental benchmarks validate the efficacy of our method.
关键词	Class imbalance learning class-incremental learning residual path
DOI	10.1109/TPAMI.2023.3275585
收录类别	SCI
语种	英语
资助项目	National Key Ramp;D Program of China ; Natural Science Foundation of China(NSFC)[2018AAA0102402] ; Natural Science Foundation of China(NSFC)[61976203] ; Natural Science Foundation of China(NSFC)[62276246] ; National Postdoctoral Program for Innovative Talents[U19B2036] ; [BX20220310]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:001068816800018
出版者	IEEE COMPUTER SOC
引用统计	被引频次：1[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/21140
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Hou, Ruibing
作者单位	1.Chinese Acad Sci, Inst Comp Technol ICT, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 5100049, Peoples R China 3.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Hou, Ruibing,Chang, Hong,Ma, Bingpeng,et al. Dual Compensation Residual Networks for Class Imbalanced Learning[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,2023,45(10):11733-11752.
APA	Hou, Ruibing,Chang, Hong,Ma, Bingpeng,Shan, Shiguang,&Chen, Xilin.(2023).Dual Compensation Residual Networks for Class Imbalanced Learning.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,45(10),11733-11752.
MLA	Hou, Ruibing,et al."Dual Compensation Residual Networks for Class Imbalanced Learning".IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 45.10(2023):11733-11752.